A variety of devices and methods are known for internally sensing physiological activity and providing therapeutic stimulation for a variety of data gathering and therapeutic purposes. For example, implantable stimulation devices are known that automatically internally sense one or more of the patient's physiological parameters and selectively provide stimulation to nerve tissue as therapy, such as for epileptic conditions, pain treatment, or apnea.
Implantable medical devices which internally sense electrical signals indicative of physiological processes of the patient have typically done so by placing one or more sense electrodes in contact with the associated patient tissue. These electrodes are then connected to appropriate amplifier and/or filter circuits such that the sensed physiologically generated electrical signals are conducted and transformed into a format suitable for analysis and utilization, such as for determination of therapy delivery or clinical data analysis.
Peripheral nerves of the body, such as the vagus or phrenic nerves, offer enticing possibilities for internally measuring the activity of these or other nerves for a variety of possible uses. For example, the phrenic nerve conducts signals originating in the brain to the diaphragm to induce the diaphragm to contract resulting in an inspiration phase of the patient's cyclic respiration. Thus, the ability to directly sense activity on the phrenic nerve would provide information indicative of the inherent perceived respiration demand rather than inferential information related to respiratory demand, such as systemic CO2 concentration, or to the respiratory response, such as a minute volume measurement. Direct sensing of phrenic nerve activity would also provide the ability to diagnose a central sleep apnea (CSA) condition by directly observing the lack of or reduced phrenic nerve activity.
However, such direct nerve sensing, particularly on a chronic or long-term basis, has been inhibited by several factors. A major impediment to direct nerve sensing in the body is accurately discriminating the nerve signals from background electrical signals also present in the patient's body. Nerve activity, such as on the phrenic nerve, typically is of microvolt to fractional microvolt in amplitude and of approximately 300 to 10 kilohertz in frequency. Background “noise”, also present within the patient's body adjacent nerves of interest, is typically several orders of magnitude greater, e.g. in the millivolt range, and of comparable frequency spectra. This background noise can arise from the patient's cardiac activity, muscular activity, and conducted electrical signals induced from background electromagnetic energy, such as electrical line supply at 60/50 Hz.
As the background signal noise can be thousands of times greater in amplitude and covering a comparable frequency range to the signal of interest, e.g. nerve activity, it is a significant technical challenge to accurately isolate the nerve activity of interest from the background noise. It is known to surgically expose nerve tissue; distance the nerve from other patient tissue thus isolating a portion of the nerve from conducted noise; attach sensing electrodes to the nerve; and sense signals directly from the nerve with appropriate amplifier and filter circuits. However, such open, invasive procedures are obviously not practical on a long-term basis.
Methods and procedures are known which might be adapted for at least somewhat effectively isolating a nerve signal from background noise, however they are typically highly demanding of processor speed and power. For example, a sensed signal can be digitized and processed with a variety of digital signal processing (DSP) algorithms. However, such DSP algorithms typically are far too demanding of processor capability and power consumption than is admissible within the available processing bandwidth and limited battery capacity of an implantable device. Filtering offers limited assistance as the frequency spectra of nerve signals is comparable to that of the background noise.
Thus it will be appreciated that there is a demand for accurate reliable direct sensing of nerve activity which is suitable for long-term chronic use, such as by a battery powered implantable medical device. There is also a need for a device which can accurately sense low level signals in an environment with relatively high amplitude noise of comparable frequency spectra and provide these signals in a manner for effective use which does not overburden the available processing bandwidth or power consumption available to a battery powered implantable medical device. It would be further advantageous to provide a sensing system which facilitates implantation by the clinician and the flexibility for application to multiple nerves.